EFFECTS OF WEATHERING IN THE ROCK AND ROCK MASS PROPERTIES AND THE INFLUENCE OF SALTS IN THE COASTAL ROADCUTS IN SAINT VINCENT AND DOMINICA

XSA A. CABRIA March, 2015

SUPERVISORS: Assoc. Prof. H.R.G.K. Hack Prof.V.G. Jetten

Thesis submitted to the Faculty of Geo-Information Science and Earth Observation of the University of Twente in partial fulfilment of the requirements for the degree of Master of Science in Geo-information Science and Earth Observation. Specialization: Applied Earth Sciences SUPERVISORS: Assoc. Prof. H.R.G.K. Hack Prof.V.G. Jetten THESIS ASSESSMENT BOARD: Prof. Dr. F.D. van der Meer (Chair) Dr. Ir. S. (Siefko) Slob (External Examiner, Witteveen and Bos, Engineering Consultancy Firm)

EFFECTS OF WEATHERING IN THE ROCK AND ROCK MASS PROPERTIES AND THE INFLUENCE OF SALTS IN THE COASTAL ROADCUTS IN SAINT VINCENT AND DOMINICA

XSA A. CABRIA Enschede, The Netherlands, March, 2015

DISCLAIMER This document describes work undertaken as part of a programme of study at the Faculty of Geo-Information Science and Earth Observation of the University of Twente. All views and opinions expressed therein remain the sole responsibility of the author, and do not necessarily represent those of the Faculty.

i

ABSTRACT

Weathering in man-made slopes, such as road cuts, is said to be accelerated by stress relief and the method of excavation and this will influence the rate at which the rock mass properties deteriorate in the engineering lifetime of the slope. It is shown in this study that the response of different volcanic rock and the associated volcanoclastic rocks respond to weathering differently, but still reflective of the weathering susceptibility dictated by their mineral composition. In general, the rocks show consistent reduction of the intact rock strength and conditions of the discontinuity. The discontinuity spacing however increases from moderate weathering degree to complete weathering degree. Indicators of salt weathering are observed in some of the rock masses exposed in the coastal areas of Saint Vincent and Dominica. The matrix of the lahar deposits show indications of granular disintegration while andesite and dacite clasts exhibit disintegration through scaling. Honeycomb structures and the tafoni are seen in the andesite lava flow unit. In the ignimbrites and block-and-ash flow deposits, the presence of the hardened surface can also be attributed to the influence of salts. The estimated rate of cavity development in andesites is at 2.5 cm/year while the estimated rate of retreat of the matrix materials of the lahar deposits is 30 cm in 55 years.

ii

ACKNOWLEDGEMENTS

iii

TABLE OF CONTENTS 1.  INTRODUCTION .............................................................................................................................................. 9 

1.1.  Research Background ............................................................................................................................................. 9 1.2.  Research problem ................................................................................................................................................... 9 1.3.  Constraints and limitations .................................................................................................................................. 10 1.4.  Objectives .............................................................................................................................................................. 10 1.5.  Research questions ................................................................................................................................................ 11 1.6.  Thesis structure ..................................................................................................................................................... 11 

2.  literature review .................................................................................................................................................. 13 2.1.  Stress relief ............................................................................................................................................................. 13 2.2.  Weathering process ............................................................................................................................................... 13 2.2.1.  Physical or mechanical weathering ...................................................................................................... 13 2.2.2.  Chemical weathering ............................................................................................................................. 13 2.2.3.  Biological weathering ............................................................................................................................ 14 2.3.  Weathering intensity, rate and susceptibility of intact rock and rock mass ................................................... 14 2.4.  Classification of weathered intact rock material and rock mass ..................................................................... 15 2.4.1.  The British Standards: BS5930:1981 and BS5930:1999 ................................................................... 15 2.4.2.  ISO 14689-1 ............................................................................................................................................ 19 2.5.  Weathering effects on the geotechnical properties of intact rock and rock masses .................................... 19 2.5.1.  Response of various rock types to weathering .................................................................................. 19 2.5.2.  Weathering effects on discontinuities ................................................................................................. 20 2.5.3.  Changes in the strength parameters due to weathering .................................................................. 20 2.6.  Weathering-time relation in rock mass classification ....................................................................................... 20 2.7.  Influence of salt ..................................................................................................................................................... 21 2.7.1.  Mechanism of salt weathering .............................................................................................................. 21 2.7.2.  Rate of salt weathering .......................................................................................................................... 22 2.7.3.  Factors governing salt weathering ....................................................................................................... 22 2.7.4.  Cementing effect of salt ........................................................................................................................ 23 

3.  STUDY area ....................................................................................................................................................... 24 3.1.  Location , topography and climate ..................................................................................................................... 24 3.2.  Geology .................................................................................................................................................................. 24 3.2.1.  Saint Vincent ........................................................................................................................................... 24 3.2.2.  Dominica ................................................................................................................................................. 25 

4.  METHODOLOGY .......................................................................................................................................... 26 4.1.  General Approach................................................................................................................................................. 26 4.2.  Desk study ............................................................................................................................................................. 26 4.3.  Field survey ............................................................................................................................................................ 27 4.3.1.  Defining and naming geotechnical units (GU) ................................................................................. 27 4.3.2.  Assigning rock mass weathering grade ............................................................................................... 27 4.3.3.  SSPC parameters for weathering (WE) and method of excavation (ME) .................................... 27 4.3.4.  Description of rock material and rock mass properties ................................................................... 28 4.3.5.  Sampling .................................................................................................................................................. 30 4.4.  Laboratory Analysis .............................................................................................................................................. 30 4.4.1.  Grain size separation and analyses ...................................................................................................... 30 4.4.2.  Clay mineralogy ...................................................................................................................................... 30 4.4.3.  Water extractable salts ........................................................................................................................... 30 

iv

4.5.  Data analysis .......................................................................................................................................................... 31 4.5.1.  Reference Intact Rock Strength (RIRS) .............................................................................................. 31 4.5.2.  Overall discontinuity spacing (SPA) and Reference Overall Discontinuity Spacing (RSPA) .... 32 4.5.3.  Condition of discontinuities .................................................................................................................. 32 4.5.4.  Reference rock mass friction angle (RFRI) and cohesion (RCOH) .............................................. 33 4.5.5.  Slope rock mass properties (SRM) and slope geometry ................................................................... 33 4.5.6.  Slope stability probability ...................................................................................................................... 34 

5.  Slope Characterization ....................................................................................................................................... 36 5.1.  Introduction .......................................................................................................................................................... 36 5.1.1.  Exposure SV1 in Saint Vincent ........................................................................................................... 36 5.1.2.  GU SV1A ................................................................................................................................................. 36 5.1.3.  SV1B ......................................................................................................................................................... 37 5.1.4.  GU SV1C ................................................................................................................................................. 38 5.1.5.  GU SV1D ................................................................................................................................................ 38 5.2.  Exposure D10 in Dominica ................................................................................................................................ 39 5.2.1.  GU D10A ................................................................................................................................................ 39 

6.  results and discussion ......................................................................................................................................... 42 6.1.  Changes in the intact rock and rock mass properties with weathering degree ............................................. 42 6.1.1.  Intact Rock Strength (IRS) .................................................................................................................... 42 6.1.2.  Spacing of Discontinuities..................................................................................................................... 44 6.1.3.  Condition of Discontinuities ................................................................................................................ 47 6.2.  Rock mass friction angle (FRI) and cohesion (COH) ..................................................................................... 49 6.3.  Weathering intensity rate ..................................................................................................................................... 50 6.4.  Weathering degree of GUs in the slope stability classes ................................................................................ 52 6.5.  Summary ................................................................................................................................................................ 53 

7.  Influence of salts in the rock masses along coastal roads in Saint Vincent and Dominica .................... 55 7.1.  Introduction .......................................................................................................................................................... 55 7.2.  Characteristics of rock masses exposed to sea spray ....................................................................................... 55 7.2.1.  Andesites .................................................................................................................................................. 55 7.2.2.  Lahar deposits ......................................................................................................................................... 57 7.2.3.  Block-and-ash flow deposits and ignimbrites .................................................................................... 59 7.3.  Results of water extractable salt experiment ..................................................................................................... 60 7.4.  Discussion ............................................................................................................................................................. 61 7.4.1.  Influence of rock properties ................................................................................................................. 61 7.4.2.  Influence of distance from the coast, presence of buffer and slope direction ............................ 63 7.4.3.  Estimates on the rate of development of salt weathering associated structures ......................... 63 7.5.  Implications of salt influence on the engineering properties of the affected rock masses ......................... 64 

Summary ....................................................................................................................................................................... 65 8.  Conclusions and recommendations ................................................................................................................ 67 References .................................................................................................................................................................... 69 9.  Appendix .............................................................................................................................................................. 76 

9.1.  General geologic maps of Saint Vincent and Dominica ................................................................................. 76 9.2.  Slope description and characterization .............................................................................................................. 77 9.2.1.  Exposures in Saint Vincent ................................................................................................................... 78 9.2.2.  Exposure in Dominica ........................................................................................................................... 83 9.3.  Scatter plots of rock properties with SSPC WE of individual GUs ............................................................. 89 9.4.  Scatter plots of rock properties vs. time of exposure ..................................................................................... 92 

v

vi

LIST OF FIGURES Figure 1. General stability and weathering characteristics of common rock‐forming minerals ........... 15 Figure 2. Description of large‐scale ....................................................................................................... 29 Figure 3. The set‐up for the water extractable salts experiment. ......................................................... 31 Figure 4. Discontinuity spacing factors (from Taylor (1980) in Hack (1998)) ........................................ 32 Figure 5. Probability for orientation dependent slope stability ............................................................ 34 Figure 6. Probability of orientation independent slope stability. ......................................................... 35 Figure 7. Exposure SV1 along the Windward Highway in Saint Vincent. .............................................. 36 Figure 8.  Highly weathered basalt in GU SV1A. .................................................................................... 37 Figure 9. Highly weathered pyroclastic flow deposit in GU SV1B. ........................................................ 37 Figure 10. Block‐and‐ash flow (BAF) deposits in GU SV1C. ................................................................... 38 Figure 11. Tuff overlying the other GUs in GU SV1D. ............................................................................ 39 Figure 12.  Exposure D10 . ..................................................................................................................... 39 Figure 13. Discontinuities in GU D1A..................................................................................................... 40 Figure 14. Columnar blocks ................................................................................................................... 40 Figure 15.  Smaller blocks formed by discontinuities in GUs D10D and D10E ...................................... 41 Figure 16.  Average Intact Rock Strength (Ave. IRS) vs. degree of weathering. .................................... 43 Figure 17.  Average discontinuity spacing (Ave. DS) vs. degree of weathering.. .................................. 44 Figure 18. Average SPA (Ave. SPA) vs. degree of weathering.   ............................................................ 45 Figure 19. Apertures of discontinuities in highly weathered rocks. ...................................................... 45 Figure 20. Joints becoming less evident with increasing degree of weathering. .................................. 46 Figure 21.  Discontinuities in completely weathered exposure and highly weathered Vlcs GUs ......... 46 Figure 22. Unloading joints resulting from combined weathering and stress relief ............................. 47 Figure 23. Average TC (Ave. TC) vs. degree of weathering.. ................................................................. 48 Figure 24. Average CD (Ave. CD) vs. degree of weathering. ................................................................. 48 Figure 25. "Flowing" Im in tuff beds. ..................................................................................................... 48 Figure 26. Average Friction Angle (Ave vs. degree of weathering.. ...................................................... 49 Figure 27. Average Cohesion (Ave. Cohesion) vs. degree of weathering.  . .......................................... 49 Figure 28. Degree of weathering of the GUs in the OIS stability classes. ............................................. 53 Figure 29. Degree of weathering of GUs in the ODS‐sliding criterion classes. ...................................... 53 Figure 30. Degree of weathering of GUs in the ODS‐toppling criterion classes. ................................... 53 Figure 31. Indicators of salt weathering in Exposure SV10. .................................................................. 55 Figure 32.  Cavities probably caused by salt weathering in the andesites. ........................................... 56 Figure 33. Honeycomb structures in exposure CM2. ............................................................................ 56 Figure 34. Examples of lava flow exposures in Saint Vincent and Dominica that do not exhibit visible indications of salt weathering ............................................................................................................... 57 Figure 35. Exposures of lahar deposits in the southeastern side of Dominica ..................................... 57 Figure 36. Indicators of salt weathering in the lahar deposits. ............................................................. 58 Figure 37. Hardened surfaces probably due to salt influence on  the ignimbrite and BAF deposits .... 59 Figure 38.  BAF exposure in a quarry in Penville,  Dominica ................................................................. 60 Figure 39. Estimated elevation (m) vs. salt concentration (ppm). ........................................................ 61 Figure 40. Estimating rate of salt weathering.   .................................................................................... 64 

vii

Figure 41. Incipient development of new discontinuity sets in Exposure SV10 due to salt weathering ............................................................................................................................................................... 65 Figure 42. General geologic maps available from literature. (a) Saint Vincent (b) Dominica ............... 76 Figure 43. Location of investigated exposures in (a) Saint Vincent; (b) Dominica ................................ 77 Figure 44. IRS vs. SSPC WE. ................................................................................................................... 89 Figure 45. .SPA vs. SSPC WE. ................................................................................................................. 89 Figure 46.  TC vs. WE. ............................................................................................................................ 90 Figure 47.  SPA vs. WE ........................................................................................................................... 90 Figure 48.  TC vs. WE. . .......................................................................................................................... 91 Figure 49. CD vs. WE . ............................................................................................................................ 91 Figure 50. CD vs. WE. ............................................................................................................................. 92 Figure 51. Relationship of time with SPA and IRS.   .............................................................................. 92 Figure 52. SFRIC vs. exposure time ....................................................................................................... 93 

viii

LIST OF TABLES Table 1.  The SSPC correction values for the method of excavation        21 

Table 2. The BS5930:1999 classes for strength of rock material          23 

Table 3. Abbreviations and notations used in Figures 1‐3, 9‐10, 12‐16          31 

Table 4. The apparent rate of weathering expressed as reduction in the rock properties                                       obtained subtracting the values in RRM  divided by a logarithmic function of time  32 

Table 5. Estimated rate of reduction in the rock properties of the GUs in Exposure SV1 using     the reference slope approach and the RRM‐SRM concept        32 

Table 6. Summary of results of the probability of OIS and ODS stability classification    46 

Table 7. Concentration of water extractable salts in samples from various exposures    with different estimated elevation above msl             54  Table 8. Computed tensile strength of the investigated rocks and values from literature    55  

Table 9. Pressure produced by salt processes (after Goudie & Viles, 1997)      56 

Table 10. Concentration of water extractable salts in samples from various exposures with  different estimated elevation above msl             64    Table 11. Computed tensile strength of the investigated rocks and values from literature   66 

Table 12. Pressure produced by salt processes (after Goudie & Viles, 1997)  66

9

1. INTRODUCTION

1.1. Research Background The weathering phase is a preparatory stage of slope denudation wherein significant modification in the engineering properties of intact rocks and the rock masses occur (Dearman, 1974). These changes are dependent on the intrinsic properties of the rock materials and on the prevailing environmental conditions. (Hencher & McNicholl, 1995; ANON., 1995; Price, 1995; Hack et al., 2003; Huisman, 2006; Tating et al., 2013). Newly exposed rock masses resulting from engineering works, also referred to as man-made slopes, are subject to accelerated deterioration due to the release of confining pressure or stress relief, and general disruption of its quasi equilibrium state that leads to intensified weathering right after excavation (Hack & Price, 1997; Nicholson, 1997; Niini et al., 2001; Huisman, 2006). Despite the general knowledge that stress relief and weathering inevitably lead to rock decay and eventual slope failure, it is still poorly integrated in the formulation of geotechnical models and oftentimes overlooked as a cause of repeated failure (Hencher & McNicholl, 1995; Lee & Hencher, 2009). The limited understanding and appreciation, as well as the limited quantitative information, on stress relief and weathering are the main reasons why these processes are oftentimes neglected or given little consideration in slope designs (Tating et al., 2013). The effects of stress and weathering are highly influenced by the composition of the rocks and on the prevailing environmental conditions. To gain more understanding on this field of study, this research focuses on the changes in the engineering properties due to stress relief and weathering of rock masses along in roadcuts in the Saint Vincent and Dominica. Both of these islands are underlain by young volcanic and volcanoclastic rocks and are located in a warm, humid environment. In addition, the main roads interconnecting most of the towns mainly traverse coastal areas. In a typical rocky coast profile, most of the coastal roadcuts are located in the sea spray zone and thus the rock masses are likely influenced by marine salts carried by sea sprays (Mottershead, 2013). Previous studies have shown the important role of salts in landscape development ( Johannessen et al., 1982; Rodriguez-Navarro & Doehne, 1999; Hampton & Griggs, 2004; Lawrence et al., 2013) and their deleterious effects on rocks used as building and construction materials (Benito et al., 1993; Benavente et al., 2007; Zedef et al., 2007; Kamh, 2011). Therefore, the rock masses that are exposed to sea sprays are also investigated for any indications of salt influence and its implication in their engineering properties.

1.2. Research problem Slope stability problems related to weathering results from the failure to recognize during site investigation and to consider in the design phase that a particular slope consists of zones with different degree of weathering and hence with varying engineering properties that also change as the rock mass further weathers through time (Hencher & McNicholl, 1995). Previous studies have shown the relationship of weathering with the degradation of the geotechnical properties of rock masses (e.g. Gupta & Rao, 2000;Tuǧrul, 2004;Arıkan, 2007). These studies have generated a lot of information collected from extensive laboratory analyses. However, these laboratory tests are very expensive to conduct and the collection and transport of samples to ensure that these meet the criteria (e.g. enough volume, representativeness and whether disturbed of undisturbed etc...) of each test is quite challenging. Field observation and in situ assessments combined with empirical models are sufficient in the initial stages of site investigation.

10

Landslides in natural slopes and failures in man-made slopes are common in Saint Vincent and Dominica. More than conducting slope stability assessment in these islands, this research uses the slope stability parameters to determine the weathering dependent changes in the rock properties control the strength of the studied rock masses This is done using the empirically-derived weathering reduction parameter (WE) introduced by Hack (1998) and imbedded in the Slope Stability Probability Classification (SSPC). This classification allows for quantitative evaluation of changes in the intact rock and rock mass engineering properties from their undisturbed states to their current conditions. Using the same principle, the future values of these properties are also be estimated and incorporated to determine future slope stability scenarios. In both islands, some of the road cuts are located very close to the sea and thus, are exposed to the influence of marine salts carried by sea sprays (Mottershead, 1989). Some of the features that are known to be associated with salt influence, such as honeycomb structures, tafone, scaling and pitting or rock surfaces, are also present in some of the exposures in the coastal areas in both islands. Previous studies have shown that salts have negative impact on the stability of coastal cliffs (Hampton et al., 2004; Lawrence et al., 2013). The presence of such features indicates that the rock masses are affected by salts and this is likely to have implications on their engineering properties.

1.3. Constraints and limitations The climatic and geologic settings of these two islands however provide constraints and limitations that can potentially become sources of data uncertainty.

The extensive chemical weathering typical in young, volcanic terrains in tropical regions (Aristizábal et al., 2005; Jain, 2014), high erosion rate (Radet al., 2013) and the thick vegetation cover resulted to gaps in the observed weathering grades for most of the rocks.

Because volcanism in these islands has been very active in the recent geologic times (Smith et al., 2013), most exposures consist of various facies of pyroclastic materials. These are highly heterogeneous and are hardly fitting in the existing weathering rock mass classification methods. Price (1995) stressed that before relating the weathering grade to the measured engineering parameters, it is important to note that a systematic description of the existing weathering conditions of the rocks is necessary. Furthermore, the heterogeneity also causes deviation from the general trends of the weathering-induced changes in the engineering properties of the rocks. The ubiquity of the pyroclastic materials makes the sampling points biased to this rock types over the others.

Especially in the case of Dominica, the good rock exposures are located in high, steep slopes where rock fall is regularly occurring. This limits the ease and thus, accuracy in the level of observation.

1.4. Objectives The general objective of this research is to determine the effect of weathering on the geotechnical properties of rock masses and the possible influence of salts in the rock masses exposed in the coastal roadcuts in Saint Vincent and Dominica.

11

The specific objectives are:

To determine the engineering parameters of rock masses in selected slopes using the Slope Stability Probability Classification (SSPC) method and relate this to the degree of weathering of the rock masses

To describe the weathering-time relationship from trends exhibited by the results of the rock mass classification to the length of year the slope has been exposed

To determine influence of salts and infer on the implications in the engineering properties of the rocks and rock masses in the coastal roadcuts

1.5. Research questions What is the applicability of the weathering classification recommended in BS5930 to the

rock masses in Saint Vincent and Dominica? How are the values of the rock properties changing with increasing degree of weathering? What are the factors influencing the weathering intensity rate of the rock properties in

the studied rock masses? How are the weathering classes distributed among the SSPC stability classes? What are the distinct features in the exposures influenced by sea sprays and what do

these indicate? What are the implications of salt influence on the engineering properties of the affected

rock masses?

1.6. Thesis structure Chapter 1- Introduction: Provides the research background and the statement of the research problem, the objectives, and the questions addressed in the research. Chapter 2 - Literature review: Provides a discussion of results and facts obtained from previous works related to the weathering process and its relationship with the engineering properties of the intact rock and rock masses, including presence of salt and salt weathering processes. Chapter 3 - Description of the study area: Describes the topography, climate, location, and the general geology of Saint Vincent and Dominica. Chapter 4 - Methodology: Describes the general approach of the research, the classification schemes used for weathering and strength, the SSPC parameter values, the laboratory procedures and the equations The equations are used calculating the geotechnical parameters and determining the slope stability (mostly from the SSPC method) and the weathering rates used and followed in the research. Chapter 5 - Slope Descriptions and Characterizations: Presents samples of field characterization of slopes. The complete description included in Appendix 1. Chapter 6 - Results and Discussion: Includes presentation and discussion of the results of the data analysis on the effects of weathering on the engineering properties of rock masses, weathering rate and stability probability classification (SSPC).

12

Chapter 7 - Discussion on influence of salts: Includes the field observation in exposures with indicators of salt influence observed in the field and their implication on the engineering properties of the affected rock masses. Chapter 8 - Conclusions and recommendations: Answers to the research questions and recommendations for future research

13

2. LITERATURE REVIEW

2.1. Stress relief The change in the stress regime following the removal of confining pressure is one of the most dominant deterioration mechanisms affecting man-made slopes upon excavation (Huisman, 2006; Tating et al., 2014). Stress relief can result to the opening of existing cracks and development of new ones within the intact rocks as illustrated in Figure 1 (Hack & Price, 1997; Price, 2009). As a result, the rock mass becomes further exposed to weathering as increased discontinuities allow more water to ingress and plant roots to reach a larger area of the rock mass (Price, 1995). Lateral stress relief upon excavation in over-consolidated clay materials cause outward movement (Waltham, 2002). Secondary stress relief is in the form of unloading after erosion where a set of discontinuities develops parallel to the ground surface (Gamon, 1983) or parallel to the erosion surface (Price, 1995).

2.2. Weathering process Weathering is the in situ breakdown of intact rock and rock masses due to physical and chemical processes under the influence of atmospheric and hydrospheric factors (Hack, 2006) and this implies decay and change in state from an original condition to a new one (Price, 2009). It is an irreversible response of soil and rock materials to their natural or artificial exposure to the near surface engineering environment ( Price, 1995). The changes resulting from weathering is a product of the interplay of structure and type of parent material, groundwater, climate, time, topography and organisms (Dearman, 1974). Through time, weathering can also be influenced by changes in land use and in the quality of the percolating groundwater as an effect of chemicals from sewage, fertilizers etc… (Hack & Price, 1997).

2.2.1. Physical or mechanical weathering Physical or mechanical weathering is the disintegration of a rock material into smaller pieces without any change in the original property of the rock. It usually results from temperature and pressure changes. The main mechanisms for this type of weathering are wedging, exfoliation and abrasion. In the tropics, repeated drying and wetting results to heaving and eventual break down of the rocks. Exfoliation occurs when rock layers break apart due to the removal of confining pressure such as when slopes are excavated (Huisman et al., 2011) or eroded (Gamon, 1983). Abrasion is the physical grinding of rock fragments either by action of water or air. Several mechanical weathering processes, such as salt weathering (more details in Section 2.3), involve the growth of a solid substance along the confining space of a pore exerting tensile stress along the pore walls and which exceeds the tensile strength of the pore leading to splitting and eventual disintegration of the rocks (Wellman & Wilson, 1965; Matsukura & Matsuoka, 1996).

2.2.2. Chemical weathering This process involves the formation of new minerals (clays and salts) when minerals react with water. This process is more favoured in warm, damp, climates. The most common processes of chemical weathering are dissolution, hydrolysis and oxidation. Dissolution mainly occurs when certain minerals are dissolved by acidic solutions and the most common example is the formation of caves in limestones due

14

to the dissolution of calcite by carbonic acid. Hydrolysis occurs when pure water ionizes and reacts with silicate minerals and it is assumed that the original mineral is transformed to a totally new mineral. Oxidation or rusting involves the combination of certain metals with oxygen allowing electron transfer leading to the formation of crumbly and weak rocks(Colman & Dethier, 1986).

2.2.3. Biological weathering Biological weathering encompasses weathering caused by plants, animals and microbes. For examples, some organisms release acidic and chelating compounds as well as inorganic nutrients that enhance chemical weathering. A species of lichens was found to cause incipient weathering of basalts by glass dissolution and precipitation of secondary carbonates and oxides (Meunier et al., 2014). Furthermore, microorganisms can oxidize organic or mineral compounds that they use as source of energy for their growth and reproduction(Lerman & Meybeck, 1988). The ability of large plants species like trees to thrive in rocky slopes that their roots and the associated mircroorganisms can potentially induce mineral weathering (Boyle& Voigt 1973).

2.3. Weathering intensity, rate and susceptibility of intact rock and rock mass Weathering intensity refers to the degree of decomposition of intact rock and rock masses (Huisman, 2006). For rock mass mass classification purposes, standardized weathering classification schemes such as the BS5930 1981/1999) are commonly used. Other methods of describing weathering intensity are through measurement of mechanical index properties (Ceryan et al., 2007) of by using chemical indices (Gupta & Rao, 2001). The weathering intensity rate is the amount of change in the weathering intensity, or just a certain amount of change per unit time. Huisman (2006) presented studies suggesting that weathering intensity rates are decreasing with time as the rock mass attains equilibrium with its surroundings. Weathering susceptibility in this context is the susceptibility to weathering of the rock or soil mass at the end of the slopes' engineering life span (Price, 2009). Figure 1 shows the general stability and weathering characteristics of common rock-forming minerals. Although not shown in the figure, gypsum, weather easily and its effect on the rock mass is observable within a short span of time after excavation. However, for rock masses with relatively resistant components, the susceptibility to weathering can be assessed based on exposures of the same rock type with known excavation date. This concept is important in slope stability because the changes in the geotechnical properties of the rock mass due to weathering can cause failure to occur even before slopes reach their designed lifetime. The accuracy on the estimation of weathering susceptibility is highly dependent on the experience of the worker and also on the rock mass factors such as regularity of weathering over the years, quantity of exposures in the area, exposure time, number of degree of rock mass weathering and the homogeneity of the rock mass (Hack, 1998).

15

Fastest

Weathering

Mineral Least

Stable

Halite Calcite Olivine Ca-plagioclase Pyroxene Amphibole Na-plagioclase Biotite Orthoclase (K-feldspar) Muscovite Clays (various types) Quartz Gibbsite (Al-hydroxide)Slowest Weathering

Hematite (Fe-oxide) Most Stable

Figure 1. General stability and weathering characteristics of common rock-forming minerals (modified from http://www.columbia.edu/~vjd1/weathering.htm, viewed 16 February 2015)

2.4. Classification of weathered intact rock material and rock mass The main purpose of having a mass classification scheme for engineering purposes, is “to provide short-hand descriptions for zones of rock of particular qualities to which can be assigned engineering characteristics within a single project” Anon.(1995). It is a means to transfer experience from one situation to another but keeping in mind that the effects of weathering varies from every rock type. A comprehensive summary and comparison of the existing weathering schemes used and recommended by researchers from 1955 to 1982 and from 1955 to 1995 as part of the effort to standardise characterization of weathered rocks and their engineering properties, were made by Gamon (1983) and Anon.(1995), respectively. The state of weathering is characterized by the degree of discoloration, decomposition and disintegration. In both papers, the authors agree that there is no single classification scheme that can encompass the complexity of weathering nor can classification be made based on a single material attribute. Hencher & McNicholl (1995) proposed a zonal weathering classification. This can be very helpful in determining which among the other existing classification methods, e.g., Anon.(1995), is applicable for a certain zone.

2.4.1. The British Standards: BS5930:1981 and BS5930:1999 The weathering classification in the BS5930:1981 is among the commonly used rock mass classification schemes. However, many researchers regard it as over simplistic and often inappropriate (Anon., 1995). In a recent review by Hencher (2008), he commended that this scheme “doesn’t work well in practice and conflicts with other well-established classifications” and it also lacks weathering classification on intact weathered rock samples while it is supposed to be used in geotechnical logging of boreholes. A revised version of the weathering classification scheme in BS5930:1981 was incorporated in the BS5930:1999 following the points raised by (Hencher, 2008). This new version consists of five approaches that cover uniform and heterogeneous materials. In this document, it is explicitly stated that the subclasses are rather

16

broad and thus its usage should be coupled by local experience, site-specific studies and by consulting other established schemes. Hencher (2008) found the 1999 version to be compatible with other schemes. Hack (1998) proposed a comparison scheme for the application of the old and new versions (Table 1). This is with reference to the weathering factors incorporated in the Slope Stability Probability Classification (SSPC) (Chapter 2.6 and Chapter 4) which is based on the 1981 version. Based on the table, the description of the moderately weathered to completely weathered weathering grades in the BS5930:1999 are already encompassed by rock masses which are classified as moderately weathered in BS5930:1981 and the completely weathered degree of Approach 1 can be under the high weathering grade of BS5930:1981. If this classification is used, then there will a single reduction value for rock masses that are moderately weathered to highly weathered which is practically unlikely. Thus

17

Table 1 Comparison for the application of BS5930:1981 and BS5930:1999 (from Hack , 1998)

BS5930 1981 BS5930 1999

Degree Description

approach 2 Uniform materials

(moderately strong or strong rock in fresh state)

Approach 3 Heterogeneous masses

(mixture of relatively strong and weak material)

Approach 4 Material and mass

(moderately weak or weaker in fresh state)

Grade Description Zone Description (2) Class Description

I Fresh

No visible sign of rock material weathering; perhaps slight discoloration on major discontinuity surfaces.

I Fresh

Unchanged from original state 1 100 % grades I -

III A

Unweathered

Original strength, colour, fracture spacing

II Slightly weathered

Discoloration indicates weathering of rock material and discontinuity surfaces. All rock material may be discoloured by weathering.

II Slightly weathered

Slight discolouration, slight weakening 2 > 9 0% grades III

18

BS5930 1981 BS5930 1999

Degree Description

approach 2 Uniform materials

(moderately strong or strong rock in fresh state)

Approach 3 Heterogeneous masses

(mixture of relatively strong and weak material)

Approach 4 Material and mass

(moderately weak or weaker in fresh state)

Grade Description Zone Description (2) Class Description disturbed

VI Residual soil

All rock material is converted to soil. The mass structure and material fabric is destroyed. There is a large change in volume, but the soil has not been significantly transported.

VI Residual soil

Soil derived by in-situ weathering but having lost retaining original texture and fabric

6 100% grades IV - VI

E residual or reworked

Matrix with occasional altered random or apparent lithorelicts, bedding destroyed. Classed as reworked when foreign inclusions are present as a result of transportation

19

2.4.2. ISO 14689-1 This weathering classification scheme made a distinction between intact rock and rock masses shown in Table 2 (modified from Mulder et al., 2012). It adapts the BS5930:1981 weathering classification for rock mass but the notation for weathering grades was modified. In this classification, the fresh rock, which is noted as I in BS5930 is assigned a grade of 0 and the residual soil, which is noted as V in BS5930 is assigned a grade of 5. This move was highly criticized by Hencher (2008) as not very helpful and not usable in practice and that adjusting the weathering grade notation provides further confusion. He also warned against the poor practice of characterizing a small sample (e.g. core from borehole), even conducting laboratory tests, then translating the results to the whole rock mass.

Table 2. Weathering description and classification of intact rock material (modified from Tables 2 and 13 from (NEN-EN-) ISO 14689-1:2003 (modified from Mulder et al., 2012; excluded weathering classification of rock mass )

Classification Description

Fresh No visible sign of weathering/alteration of rock material Discoloured The colour of the original fresh rock material is changed and is evidence of

weathering/alteration. The degree of change from the original colour should be indicated. If the colour change is confined to particular mineral constituents, this should be mentioned.

Disintegrated The rock material is broken up by physical weathering, so that bonding between grains is lost and the rock is weathered/altered towards the condition of a soil in which the original material fabric is still intact. The rock material is friable but the mineral grains are not decomposed

Decomposed The rock material is weathered by the chemical alteration of the mineral grains to the condition of a soil in which the original material is still intact; some or all of the mineral grains are decomposed.

2.5. Weathering effects on the geotechnical properties of intact rock and rock masses

2.5.1. Response of various rock types to weathering Various rock types respond to weathering in various ways. For volcanic rocks, the reaction of water converts the volcanic glass into clay and this causes volumetric changes that would further promote physical and mechanical changes in the inter-granular structures (Yokota & Iwamatsu, 1999). Volcaniclastic rocks may generally behave like conglomerates with the matrix materials sometimes behaving as sandstones. Chigira & Sone (1991) studied the weathering profile of young sandstones and conglomerates and identified weathering zones of oxidation to dissolution through depth. The mechanical properties of the rock mass vary systematically with the change of the weathering zone. Gupta & Rao (2000) presented studies showing that in granites, the loss of strength from fresh to moderately weathered rocks reaches about 80%. For claystones, the tensile strength observed in fresh rocks is decreased by 75% in slightly weathered rocks because of the increase in microfractures. Results of petrographic analyses suggest that microfractures, pores and voids are the dominant factors that govern the strength of fresh rocks and not the mineralogy itself. Gurocak & Kilic (2005) studied the weathering effects on the properties of Miocene basalts in Turkey classified using ISRM weathering classification. Their results showed that UCS derived from Schmidt hammer tests, the compressive wave velocity and unit weight decrease while porosity and water absorption increase with increasing degree of weathering.

20

2.5.2. Weathering effects on discontinuities Discontinuities can be mechanical or integral (Hack, 2006). Mechanical discontinuities are planes of weakness such as bedding planes and joints where the shear strength is significantly lower than the surrounding rock material. Integral discontinuities have the same shear strength as the surrounding materials such that it does not affect the intact rock strength. Discontinuities are also modified by weathering (Hack&Price, 1997). After stress relief where new cracks in the intact rocks can develop and existing cracks are opened, weathering subsequently weakens the discontinuity wall and infill materials. Further weathering will cause discontinuity planes to smoothen due to the loss of their asperities. Tating et al. (2014) noted new mechanical discontinuity sets formed thus a decrease in the spacing was observed with increasing degree of weathering in massive sandstones. However, Ehlen(2002) observed that some discontinuities disappear or become less persistent in more weathered granite and attributed this to the infilling of mineral grains in joint apertures, eventually obscuring the individual joints. This may lead to inaccurate rock mass classification such that careful assessment should always be practiced.

2.5.3. Changes in the strength parameters due to weathering Weathering leads to the disruption of grain to grain bonding creating micro-fractures and new minerals. This inevitably results to modifications in the rock mass engineering properties (Gupta & Rao, 2000). These changes include decrease or loss of intact rock strength and rock mass strength, increase in their deformability and changes in the permeability depending on the nature of the rock and its stage of weathering (Hencher & McNicholl, 1995) and this usually leads to the deterioration and subsequently, slope failure (Huisman, 2006; Fan et al., 1999; Gupta & Rao, 2000; Calcaterra & Parise, 2010; Tating et al., 2013). Parameters that are highly affected by weathering as indicated by their good correlation with the degree of weathering include tensile strength (Arıkan et al.,2007), compressive strength and to some extent, elasticity modulus (Heidari et al., 2013). Index properties that change during weathering include dry density, void ratio, clay content and seismic velocity (Ceryan, 2007). These changes however occur after rocks reach certain weathering stage (Arıkan et al., 2007).

2.6. Weathering-time relation in rock mass classification Rock mass classification schemes are widely used in slope stability assessment. These include the Rock Mass Rating (RMR), the Slope Mass Rating (SMR), Q-system, among others (Nicholson, 2004). These classification systems are difficult to apply to rock masses that are of very poor quality and in heterogeneous rocks such as flysch. The geological strength index (GSI) was formulated to address this as it would place greater emphasis on basic geological observations of rock-mass characteristics, reflect the material, its structure and its geological history and would be developed specifically for the estimation of rock mass properties (Marinos et al., 2005).However, these schemes focus on the attitude of discontinuity planes and less attention is given to the weathering state of the rocks. Weathering classifications also exist (BS 5930, 1981,1999; Dearman, 1974; Ceryan et al., 2007;Arıkanet al., 2007) but these fail to treat weathering as a progressive process that affects salient geotechnical properties of the rock mass during the engineering lifetime of a cut slope. The inadequacy in considering weathering-time relation is addressed in the Slope Stability Probability Classification (SSPC) of Hack et al. (2003) and the Rockslope Deterioration Assessment (RDA) of Nicholson (2004). The SSPC is specifically designed to address slope stability while the RDA addresses shallow weathering-related erosional processes and mass movements. SSPC involves a three-step approach that take into consideration the past and future weathering and the damage resulting from the excavation method which would indicate probable failure mechanisms (Hack, 2003). A modification of the 1998 version of SSPC was made by Lindsay et al. (2001). The main modification was the introduction of rock intact strength derived from the modified Mohr-Coloumb failure criterion adapted from varying moisture content,

21

weathering state and confining pressure. The RDA addresses shallow weathering-related erosional processes and mass movements. It likewise follows a step-wise approach to slope hazard assessment involving the application of ratings to assess rockslope deterioration susceptibility, review of the nature of the likely deterioration based on rock mass type and slope morphology and providing guidance and appropriate mitigation based on the findings in the two earlier stages (Nicholson, 2004). Between SSPC and RDA, there are more published papers applying the weathering-time relation concept of SSPC than those that actually used the RDA. Tating et al. (2013) established relationship of intact rock reduction with exposure time in sandstone units and this can be used to predict further reduction within the serviceable time of slopes built in the said unit. Rijkers & Hack (2000) found that laboratory test results for friction angle, cohesion and natural slope angle correlate well with SSPC results for the pyroclastic deposits in Saba, Netherlands Antilles. Das et al. (2010) used SSPC for landslide susceptibility assessment and remarked that although it required extensive field data, it is more accurate than GIS-based quantitative modelling.

2.7. Influence of salt The influence of marine salts makes the weathering process in the coastal environment distinct (Mottershead, 2013). Numerous studies have been conducted in natural environments and in simulated laboratory conditions to study the effect of these salts on rock materials. Results commonly show that in general, the influence of generally leads to the weakening and subsequent disintegration of rock materials (Lawrence et al. 2013). Thus, the term "salt weathering". Salt weathering includes the physical process of salt crystallization with that results to rupturing of the rock thus it is an important mechanism in rock decay (Rodriguez-Navarro & Doehne (1999). It plays an important role in the development of many geomorphologic features in coastal environment (Mottershead, 1989; Goudie & Viles, 1997;Cardell et al., 2003; Rivas et al., 2003; Sunamura & Aoki, 2011; Lawrence et al.,2013;Hampton & Griggs, 2004) and in the degradation of many structures in archaeological sites (Kamh, 2011,2005; Mottershead et at.,2003; Lubelli et al., 2004; Sancho et al., 2003). Its damaging impact impact on engineering structures such as roads, highways, runways, dams and building foundations are shown by the studies of Benavente et al.(2007), Liu et al. (2014). Salt weathering can occur on a wide range of environment and most studies done in coastal environments in temperate regions (Lawrence et al., 2013) or arid regions (Wellman & Wilson, 1965; Brandmeier et al., 2011) . There is however, limited studies that were conducted in tropical areas (Wells et al., 2006; Bryan & Stephens, 1993) . The indicators of salt weathering have been collectively noted as cavernous cavities called honeycombs and tafoni depending on the scale; white salt efflorescence, contour scaling and stone surface exfoliation (Smith & McGreevy, 1988). In general, the temporal variability in salt accumulation along the coastal area is governed by episodes of high winds, high surfs and precipitation. The spatial distribution of salt is controlled by the elevation above the shoreline, the aspect and the presence of shelters or buffers (Mottershead, 2013). The transfer of marine salt from the ocean involves three stages: the salt is released from the ocean to the atmosphere, it travels laterally through the atmosphere and finally it gets deposited on a land surface (Mottershead, 2013). The amount of salt transferred to the land through this mechanism is highly influenced by wind speed as suggested by the positive correlation between wind speed and salt concentration(Lewis & Shwartz, 2004 in Mottershead, 2013).

2.7.1. Mechanism of salt weathering The most common mechanisms of salt weathering involve physical effects caused by the stress generated by crystal growth or moisture absorption by hygroscopic salts and chemical weathering resulting from the interaction of saline pore fluids and minerals (Mustoe, 2010). Other studies (e.g. Sperling and Cooke,

22

1984) also show that hydration of sodium sulfate can also induce rock deterioration but not as aggressive as the effect of crystal growth. The influence of salts to chemical weathering was explored by Mottershead et al., (2003) They attributed the formation honeycombs in sandstones as the result of accelerated weathering that takes place during chemical dissolution of grain boundaries under the influence of salts. The same remark is found in Goudie&Viles (1997) “Salts in general participate in chemical reactions, reacting with minerals and rock surfaces.” However, because more authors support the idea of salt crystallisation as salt weathering mechanism, they tend to regard salt weathering as a physical weathering process and does not involve major chemical processes such as hydration, as implied by Kirchner (1996).

2.7.2. Rate of salt weathering In terms of rate associated to salt weathering, Kamh (2011) measured a weathering rate of 0.42mm/year in sandstones used in ancient buildings in Aachen, Germany. This rate is said to be higher compared to 0.1mm/year previously measured in studies conducted in an area with similar climatic conditions. For long term weathering, salt weathering decelerates through time as suggested by the experiments of Sperling & Cooke (1984) and Wells et al. (2006) where material loss decrease after samples were subjected to a certain number of cycles. In addition, Wells et al. (2006) also showed that there is no significant difference in the weathering rate in schists between dry and wet seasons in a simulated tropical environment. Matsukura & Matsuoka (1996) considered depth to be the most reasonable measure of growth because it increases gradually (exponential function) with time compared with the other dimensions due to the possibility of coalescence of adjoining tafoni. They computed rates ranging from 0.595 mm/year to 0.0108 mm/year. Motterhead (1982) calculated and annual weathering grate of 0.6mm/year for greenschists and this was attributed to salt weathering due to the crystallization of halite and not on chemical dissolution. Exceptional erosion figures measured from surface lowering associated with coastal salt weathering is >1mm/year and a maximum of 5.25 mm over 20 years in insoluble rocks (Mottershead, 2013).

2.7.3. Factors governing salt weathering The factors influencing salt weathering come from the attributes of the material such as porosity, permeability, geochemistry and mineralogy of the rock (Mottershead, 2013; Kamh, 2011) and the properties of the salt including its composition and the concentration of the solution, viscosity, surface tension and vapour pressure. Environmental factors such as temperature, moisture content (Trenhaile, 2005), humidity and topography (Goudie & Viles, 1997) as well as, solar exposure (Mottershead, 2013) are also important. Sperling & Cooke (1985, in Kamh (2011)) found that hydration of sodium sulphate is effective in rock disintegration but significantly less effective than crystal growth (anhydrated versus hyrdrated (thenardite vs. mirabilite). The most disintegration occurs during extreme temperature and very low relative humidity. The porosity and permeability of the rocks enable the water to enter, circulate and remain within the material. A comparative study was done on the salt weathering in sandstones, limestones and trachytes, rocks making up the St. Maria church in Cologne, Germany. The sandstones have high macropore content and interconnecting micropores encouraging crystallisation in the interstices which leads to granular disintegration. The limestones have very fine, interconnected intergranular pores so crystallization occurs in the surface forming salt crusts which sometimes detach and pulling off some fragments. Trachyte has heterogenous pore system with fissures near weathered phenocrysts encouraging cracking and scaling of fragments (Goudie&Viles, 1997). In contrast, Auger (1990, in McLaren, 2001) noted that it is the rocks with lower porosities that are more prone to salt weathering. In porous rock, the pores allow solutions to move freely in and out of the rocks such that little weathering occurs.

23

For granites and probably applicable to other crystalline rocks such as basalts and andesites, salt weathering is linked with the increase in porosity resulting from the decomposition of the minerals through other weathering processes.

2.7.4. Cementing effect of salt It was previously mentioned that salt crystallization in pore spaces under certain conditions depending on the nature of salts and the material porosity does not necessarily exert enough pressure in the walls to cause disruption or eventual weathering. Instead, this can increase cohesion among the particles through cementation. McLaren (2001) found that “in highly porous rocks (including sandstone and limestone) aeolianite deposits in the spray zone tend to be better cemented, have higher levels of secondary porosity, lower primary porosity and a lower unaltered allochemical content than the same formations that are not exposed to sea spray. However, note that the cementing effect of salt was only described in rocs that are inherently containing carbonates. The processes can be different rocks with low carbonate content such as volcanic rocks.

24

3. STUDY AREA

3.1. Location , topography and climate Saint Vincent and Dominica are located in the east of the Caribbean Sea. Dominica has a land area 750 km2 while Saint Vincent has 344km2. Roseau is the capital city of Dominica and Kingstown for Saint Vincent. In both islands, these capital cities are located in the southwest part which is said to be relatively sheltered from hurricanes that regularly visit the region. The topography in both islands is typically characterized by a rugged, mountainous central part such that areas for settlements and road networks are only limited in the coastal region and in foot slopes (Anderson & Kneale, 1985). The central highland of Dominica is formed by MorneDiablotins (1421m) and other peaks with heights exceeding 1000 m. Volcaniclastic fans form most of the flat lands where settlements have been established. Similarly, the major topographic feature in Saint Vincent is a north-south trending chain of mountains from La Soufriere (1178 m) in the north to Mount St. Andrew (736m) to the south. Fifty percent (50%) of the total surface area has a slope of at least 30 and only less than twenty percent (20%) has a slope of less than 20. Deep-cut valleys and steep coastal cliffs characterize the leeward side and wider and flatter valleys in the windward side (USAID, 1991). Both islands are characterized by a tropical climate. However, there is a significant difference in terms of rainfall. Dominica is significantly wetter with average annual rainfall frequently exceeding 5000 mm in the east coast and just 1800 mm in the west side. In the highlands, the annual rainfall can reach up to 9000 mm. The wettest months are from June through October. In Saint Vincent, the wettest months are from May to October where average annual rainfall is about 3800 mm inland and about 2000 mm in the coast. Humidity follows the trend of the rainfall

3.2. Geology Dominica and Saint Vincent belong to the chain of volcanic islands forming the Lesser Antilles island arc. It is the surface manifestation of the subduction of the North American Plate beneath the Caribbean Plate that was initiated as early lower Eocene (Smith et al., 1980; Bouysse et al., 1990; Rad et al., 2013). The general geology of both islands are described below and the maps are included in Appendix 9.1.

3.2.1. Saint Vincent The geology of Saint Vincent is characterized by basalts emplaced during the early phase of volcanic activity in the island and followed by the andesites that occur as dikes, domes or central plugs in the vents of some volcanic centres. Basalts are dominant in the south while andesites are abundant in the northern part of the island. The Southeast Volcanics consist of scoraceous basalts interbedded with massive well-jointed basaltic lava flows. It is intruded by dikes and mostly overlain by fine grained yellow ash associated with the tephra ejected by the Soufriere volcano. The Grand Bonhomme Volcanic Center is interpreted as a stratovolcano with interbedded sequences of block and ash pyroclastic flow deposits, ashfall deposits, lava flows and subordinate domes. These rocks form a heavily forested landscape with inaccessible interior composed of deeply weathered lavas and volcaniclastic deposits. The MorneGaru volcanic centre is in the north of Grand Bonhomme. The rocks exposed are lava flows, undifferentiated volcaniclastics, red scoria bombs and yellow ashfall deposits. In the southeast and northern part of the island are poorly consolidated sequence of clast-supported, pumice lapilli airfall, scoria bombs and ash overlying old lava flows. The abundant scoria bombs that fell close to these centres formed thick and sometimes welded deposits. Ash and small projectiles deposited further from the vents produced discrete beds. Spatter cones are also exposed in the northern part of the

25

island consist of a thick sequence (>20 m) of interbedded grey lapilli-sized ash and red scoria overlain by yellow ash. The red scoria clasts are composed of olivine microphyric basalts but also contain angular basaltic-andesite. The Soufriere stratovolcano occupies the northern half of the island. It is the most active volcano in the Antilles arc. Its last five major eruptions occurred in 1718, 1812, 1902, 1971 and 1979 where basaltic lava domes were extruded in the crater area followed by a phreatomagmatic explosion that produced pyroclastic flows. Other major volcanic centers were identified but these have already become extinct (Heath et al., 1998). It is deemed that the volcanic activity in Saint Vincent is younger than the other islands. There were no deposits older than 2.8 Ma, though this may also indicate incomplete sampling.

3.2.2. Dominica Dominica is underlain by sub-aerial lava flows and pyroclastics with minor Pleistocene to Holocene uplifted conglomerates and corals in the west coast of the island (Christian, 2012). A comprehensive K-Ar and carbon dating of the volcanic rocks resulted to the subdivision of the rocks into four units (Smith et al., 2013). The Upper Miocene dominated by mafic volcanism and make up the eastern part of the island. The Upper Pliocene to Lower Pleistocene unit forms two major stratovolcanoes (proto- Morne Diablotins and Cochrane-Mahaut) and two smaller Morne Concrod and Morne Bois) stratovolcanoes located in the eastern flank of Mount Diablotins. Lower to Upper Pleistocene forms two volcanic centers, Proto- Morne aux Diables in the north and Foundland in the south, and is the least extensive unit in the island. The Upper Pleistocene – Holocene is composed of seven volcanic centers in the island which marked the renewed volcanism producing andesites and dacites.

26

4. METHODOLOGY

4.1. General Approach As indicated in the list of objectives, the purpose of the research is not to conduct slope stability assessment per se but rather to use the slope stability parameters to show how the initial stress relief and weathering are affecting the geotechnical properties, and consequently the stability, of the studied rock masses. The data obtained from the various slopes that are made up of different lithologies, with different weathering degree and with different number of exposure years, are systematically recorded and manipulated using empirical equations in the Slope Stability Probability Classification (SSPC) system (Hack, 1998). Because the degree of weathering is incorporated in the calibration of the SSPC system, it is still expected to reflect differences in the calculated stability of geotechnical units with broadly uniform conditions but with different degree of weathering (e.g. exposure of one lithology with varying weathering degree). However, the difference between the rocks types and the climatic condition in the Caribbean islands and that of the area in Spain where the model was calibrated from, may give unexpected results. These are discussed in the subsequent chapters. Defining the relationship of weathering with time can be done in two ways as long as the exposure time is known. One is using the concept of Reference Rock Mass (RRM) and Slope Rock Mass (SRM) of the SSPC system and the other is to compare actual measurements from a particular geotechnical unit exposed in a certain year with those of its counterpart in a newly exposed slope (reference slope). Unfortunately, for this research, there is only one reference slope identified during the fieldwork (further discussed in Chapter 6). In order to describe the influence of salts, the exposures along the coast were compared with the exposures found in the interior (roads traversing the forested highlands) parts of the islands. Some features that are commonly observed in areas affected by salts documented elsewhere were identified in the coastal exposures in both islands. The trend in the concentrations of salts in the samples with respect to the distance from the coast strengthens the hypothesis that these features indicate the influence of marine salts. The mechanisms of how salts are affecting the rock masses depending on the type of salt influence indicator observed are inferred using literature data. These secondary data were obtained from long term monitoring, e.g. 7 years in Mottershead (1989) of salt weathering in the natural environment and from simulations in specially designed laboratory climate rooms. Both methods are not doable in the given timeframe of this research. Nevertheless, by combining these data with the field data and limited laboratory analysis results obtained in this research, possible implications to the engineering properties of the affected rock masses can be inferred.

4.2. Desk study The east Caribbean region, where Saint Vincent and Dominica are located, has been a subject of numerous projects on management and mitigation of natural hazards relating to hurricanes and volcanic activities. The reports generated from these studies contain a considerable amount of secondary information. A geological report is available for Dominica Smith et al., 2013) and a fair description of the geology of Saint Vincent can be found on the website of the Seismic Research Centre of the University of West Indies (http://www.uwiseismic.com/General.aspx?id=66). Sources of information on the geotechnical properties of the rocks in the islands are very limited. Most of the reports on geotechnical testing are for particular engineering projects, which mostly cover a limited area and contain few details.

27

4.3. Field survey A reconnaissance survey was conducted to identify the location of particular slopes to be investigated. Among the considerations in selecting these locations include the representativeness in terms of lithology and weathering degree and the accessibility of the exposure. The distance to the base camp was also a factor given the limited time available for the fieldwork. Unfortunately, many good exposures are located in roadcut sections with high and steep slopes where rockfalls regularly occur and could not or only partially be assessed due to safety consideration.

4.3.1. Defining and naming geotechnical units (GU) In the exposures, units with broadly the same geotechnical characteristics including the same weathering degree and fracture patterns are grouped as one geotechnical unit (GU). In all cases, one geotechnical unit consists of one lithologic type. In most of the exposures, several GUs were delineated within a single lithologic type. For example, 4 GUs were identified in an extensive exposure of pillow lava flow based on the differences in their weathering degree and block sizes. The names of the GUs include the assigned exposure identification code and the letter corresponding to the unit. For example, GU SV1A means geotechnical unit A in exposure SV1.

4.3.2. Assigning rock mass weathering grade In the discussion of the various weathering classification schemes in Chapter 2, the two versions of the BS5930 classification systems were compared. Approach 3 appears to be fitting for the block-and-ash flow (BAF) deposits and tuff breccia rock masses. However, if the explanation of ANON. (1995) on the background of this classification is carefully considered, this cannot be used for the zoning of such deposits because these rocks did not originate from a homogenous unit but rather, these are intrinsically heterogeneous units even in their fresh state. This is where the relevance of the dominant component is considered. In clast-supported BAF, the rock mass weathering grade is based on the dominant weathering grade of the clasts but if it is matrix-supported, then the weathering grade assigned to the GU is based on that of the matrix. Based on Table 1, the weathering grade in BS5930:1981 and that of approach 3 is highly correlatable, thus the reduction values of the SSPC system yields the same result. Therefore, to be consistent with the SSPC system, BS5930:1981 was used in this research.

4.3.3. SSPC parameters for weathering (WE) and method of excavation (ME) The SSPC parameter for weathering (Table 3) is proposed by (Hack, 1996;1998) as a method of quantifying weathering. This is embodied in the SSPC (Hack, 1996;1998) where the effects of weathering on the intact rock strength and the rock mass spacing and conditions of discontinuities are related to the degree of rock mass weathering classification in BS5930:1981. These reduction factors allow prediction of changes in these rock properties as weathering progresses for a certain slope.

Degree of weathering in slope (BS5930:1981) SSPC WE

Fresh 1.00 Slightly weathered 0.95 Moderately weathered 0.90 Highly weathered 0.62 Completely weathered 0.35

Table 3. Reduction values per weathering grade by (H.R.G.K. Hack, 1996) as used by Huisman (2006)

28

The SSPC parameter for the method of excavation enables the exclusion of any influence of the method of excavation on the measured parameters. The correction values are in Table 4. According to the Ministry of Works (MOW) of both islands, there was no blasting and only excavators were used in the road sections investigated. Therefore the the ME value for all the slopes is 1.

Table 4. The SSPC correction values for the method of excavation

Method of Excavation SSPC ME

Natural / hand-made 1.00

Pneumatic hammer excavation 0.76 Pre-splitting / smooth wall blasting 0.99

Conventional blasting with result: good 0.77 open discontinuities 0.75 dislodged block 0.72 fractured intact rock 0.67 crushed intact rock 0.62

4.3.4. Description of rock material and rock mass properties The exposure characterization sheet of the SSPC is used to systematically record the field observations which mainly consist of the description of rock material and discontinuity properties. One sheet corresponds to one GU. The filled out sheets are in Appendix 1. The description of the exposures follows the format recommended in BS5930:1999. The complete slope description is included in Appendix 10.2.

4.3.4.1. Intact rock strength (IRS) The intact rock strength (IRS) was estimated through crumbling by hand and by using the geologic hammer as recommended by Hack & Huisman (2002) and by referring to the scale in the BS 5930:1999 (Table 5).

Table 5. The BS5930:1999 classes for strength of rock material

Field definition IRS estimate (MPa)

Crumbles in hand 200

For fresh and slightly weathered lava flows, rebound values were measured with an L - Type Schmidt hammer (serial no.: proceq L-9 5526) to estimate the Uniaxial Compressive Strength (UCS) as described by (Aydin, 2009). The conversion of the rebound values to MPa used the conversion method included in the operating manual published by the Schmidt Hammer manufacturer (Proceq, 2006).

29

4.3.4.2. Discontinuity spacing The discontinuity sets were defined visually. Discontinuities with the same dip/dip direction are grouped as one set. The notation for discontinuity orientation used in the SSPC sheets is dip/dip direction. For example, J1 - 60/150 means Joint set 1 with dip of 60° and dip direction of 150°. Bedding planes are denoted as B and faults as F. Following BS5930:1999, the discontinuity spacing (DS) was measured perpendicular to the discontinuities. The minimum spacing was considered. Only the mechanical discontinuities are measured in detail in this research. The DS is used in the SSPC system to get the corresponding rock mass spacing parameter (SPA). Closely associated with the DS is the persistence of a discontinuity, which determines the possibilities of relative movement. The persistence was measured along dip and along the strike.

4.3.4.3. Condition of discontinuities The conditions of discontinuities indicate the shear strength along the discontinuities. It is assumed that weathering will lead to the loss of roughness along the discontinuities and to the formation of clay infill materials. The results of these field data are included in the slope characterizations of the GUs in Chapter 5 and Appendix 10.2 a. Large-scale (Rl) and small-scale (Rs) roughness - The importance of the discontinuity surface roughness on the shear strength along the discontinuity planes depends on the stress configuration on the discontinuity plane and in the deformation characteristic of the discontinuity wall material and asperities (Hack, 1998). These are explained in detail in the SSPC System. The Rl and Rs were measured by assessing the wavelength and amplitudes of the discontinuity surface using the Figures 2a and 2b as reference. In the SSPC system, the ISRM profiles were modified for a new empirical relation consisting of a combination of tactile and visible roughness. The large-scale roughness is determined in an area larger than 20 cm x 20 cm but smaller than 1 m x 1 m. It is described in five classes namely wavy, slightly wavy, curved, slightly curved and straight. Tactile roughness is classified as rough, smooth and polished as distinguished by feeling the discontinuity surface with the fingers in an area of 20 cm x 20 cm. The small-scale roughness is described as stepped, undulating and planar.

+

Figure 2. Description of large-scale (Rl) and small-scale (Rs) roughness of discontinuities;(a) Rl is determined in 1 m x 1m area; (b) Rs is determined in 20 cm x 20 cm area of the discontinuity plane (Hack, 1998)

a b

30

b. Discontinuity infill material (Im) - The type of infill material in between discontinuity walls and whether the discontinuity walls are in contact or not during shearing have a very strong influence in the shear strength characteristics along the discontinuities. The materials are described as cemented, non-softening and softening when subjected to the influence of water, deformation or shear displacement. C. Karst (Ka) - The open cavities are known to considerably weaken the rock mass. In the study area, the open cavities encountered are not from karstifcation but rather as a result of the gradual removal of unconsolidated pumice and scoria surrounded by a more competent rock unit. In the SSPC system, the rating is either 1 or 0.92 which only rates the presence or absence regardless of size of the cavity.

4.3.5. Sampling Samples were collected for clay mineral identification, determining the indications of salt enrichment and for grain size analysis. Unfortunately, not all the weathering grades for all the rock types encountered were sampled mainly because no representative outcrop was observed or investigated. One of the major constraints in the quantity of samples brought back to the laboratory for analyses is related to logistics involving the strict regulations on transporting soil and rock materials from the study areas to ITC.

4.4. Laboratory Analysis The laboratory analysis including all the necessary preparatory works were conducted in the ITC laboratory.

4.4.1. Grain size separation and analyses Grain size separation through sieving was conducted during the sample preparation to separate the

31

  

Figure 3. The set-up for the water extractable salts experiment. (a) Solution after stirring and (b) after being left overnight to settle (photos by de Smeth, 2015)

4.5. Data analysis The calculation for the strength parameters (and slope stability probability) of the rock masses entirely follows the SSPC system. This SSPC system differs from other rock mass classification systems because it involves a three-step approach. Three rocks masses are considered, namely: the exposure rock mass (ERM) representing the conditions at the time of investigation, the theoretical fresh reference rock mass (RRM) corrected using the SSPC reduction factors of weathering (Table 2) and excavation on the values of the ERM and; the slope rock mass (SRM) where the actual stability assessment is conducted. Using empirical equations, the cohesion angle, internal friction angle and the stability probability were calculated. This approach allows inferences to be made on the weathering induced- strength reduction in the rocks making up the geotechnical units and thus can be taken into account in designing future road improvement measures.

4.5.1. Reference Intact Rock Strength (RIRS) For the intact rock strength, the adjustment formula to obtain the Reference Intact Rock Strength (RIRS) is: (1)

where IRS is the field estimate and WE is the correction for weathering. If IRS > 132 MPa, then RIRS is 132 MPa. The maximum value of 132 MPa is taken as the cut-off value "above which the influence of the IRS on the estimated slope stability is constant" (Hack, 1998).

a

b

32

4.5.2. Overall discontinuity spacing (SPA) and Reference Overall Discontinuity Spacing (RSPA) Oftentimes, the stability of rock masses is not only controlled by a single set of discontinuities but rather it is influenced by all the discontinuity sets. In order to include possible influence of all the discontinuity sets present in a rock mass, the SSPC system adapted the SPA for a maximum of three discontinuity sets developed by Taylor (1980). The SPA is given by the formula: 1 ∗ 2 ∗ 3 (2) where SPA is the spacing parameter and the factors are obtained from Figure 4.

The RSPA is given by: ∗

(3)

where RSPA is the spacing parameter for the RRM, SPA is the existing spacing parameter, WE is the correction parameter for weathering and ME (Table 3) is the correction factor for the method of excavation. In this research, there are exposures with maximum of five sets of discontinuities in which case the minimum, median and maximum spacing were considered.

4.5.3. Condition of discontinuities This parameter includes the condition of individual discontinuity set (TC), reference condition of discontinuity set (RTC), overall condition of discontinuities (CD) and reference overall condition of discontinuities (RCD). Th